Machine Learning Capstone Free Course For Beginners

Introduction

This advanced-level Machine Learning Capstone course is perfect for learners who want to polish their machine-learning skills and further advance their proficiency. The course focuses on using several Python-based libraries, including Pandas, sci-kit-learn, and Tensorflow/Keras. Participants in this course will be guided on how to use the aforementioned libraries to solve machine-learning problems. They will also be assigned projects where they can showcase their practical skills.

Furthermore, participants will learn how to develop recommender systems, analyze data sets, and train neural networks to predict course ratings.

What Will You Gain from This Course

In this machine learning course, participants will:

1. Learn how to create recommender systems in Python and differentiate machine learning algorithms.
2. Build recommendation systems by utilizing KNN, PCA, and non-negative matrix collaboration filtering.
3. Develop regression and classification models to predict course ratings.

Skills Acquired

1. Artificial Neural Network
2. Data Analysis
3. Python Programming
4. Supervised Learning
5. Unsupervised Machine Learning

Who Is This Course For

This course is designed for:

1. Learners with prior experience in machine learning or data science looking to advance their understanding and skills.
2. Aspiring data scientists who want to build machine learning expertise through practical hands-on projects.
3. Undergraduate or graduate students who want to enhance their applied machine learning skills.
4. People in technological fields who want to shift their careers to AI-focused fields.

Course Content

5 Modules – 6 Videos – 6 Quizzes – 3 Readings – 1 Peer review – 11 App items – 16 Plugins – Certificate of Completion

Machine Learning Capstone Overview

The introductory section will lay the foundation for subsequent sections and will primarily focus on the concept of recommender systems. Additionally, you will learn the basics regarding the capstone project.

This section features hands-on practice, allowing participants to perform exploratory data analysis, examine data patterns, calculate summary statistics, and create graphical representations, e.g., charts and graphs. The data sets used will be related to online courses, including information on course titles, genres, and enrollments. Furthermore, you will learn about the “bag of words” (BoW) features, which include extracting a word-count vector from course details. Utilizing these vectors, participants will finally be able to calculate course similarities.

1. 2 Videos
2. 2 Quizzes
3. 3 App Items
4. 5 Plugins

Unsupervised-Learning Based Recommender System

In the second module, learners will get to build three different course recommender systems. The first lab is based on building a course recommendation system using matrices that represent user-profiles and course genres. In this way, you will learn how to calculate interest scores for each course and recommend courses with the highest scores. The second lab guides participants on how to create recommendation systems using a course similarity matrix. The role of clustering algorithms in course recommendation systems is introduced in the third lab. The fourth lab teaches participants to implement K-Nearest Neighbors (KNN)-based collaborative filtering to predict user interests. Finally, in the last lab, you applied collaborative filtering using non-negative matrix factorization (NMF) to estimate user interests.

1. 1 Video
2. 2 Quizzes
3. 3 App Items
4. 2 Plugins

Supervised-Learning Based Recommender System

This section primarily focuses on predicting course ratings and understanding user behavior using neural networks. Similar to the previous section, you will learn content through a series of labs here. In the first lab, you will learn how to use neural networks to extract latent features of users and items and derive patterns to effectively predict course ratings. In lab 2, you will be given course interaction feature vectors as input data to run regression analysis and evaluate numerical rating scores. These rating scores will be used to predict whether a student will audit or complete a course. In the final lab, you will extract user and item embedding feature vectors, create an interaction feature vector, and predict whether a student will audit or complete a course using a classification approach instead of regression.

1. 1 Video
2. 2 Quizzes
3. 5 App Items
4. 1 Plugin

Share and Present Your Recommender Systems

Once you have learned how to build recommender systems, it is time to learn how to present them as well. This section is centered on guidelines for developing high-quality reports. Moreover, you will be given instructions on how to create a good PowerPoint presentation and save it as a PDF.

1. 2 Videos
2. 4 Plugins

Final Submission

The last section introduces participants to Streamlit and equips them with an opportunity to develop a Streamlit app that enables them to present their work projects from previous sections. You will compile your work from all the hands-on labs and submit it. Your peers will then review it and give their remarks. After making your submission, you will then review someone else’s work and assign an appropriate grade.

1. 3 Readings
2. 1 Peer Review
3. 4 Plugins

Description

To take this advanced-level course, participants need to fulfill certain prerequisites, including completing all prior courses in the IBM Machine Learning Professional Certificate program. Building on the foundation laid by these previous courses, this course introduces learners to recommender systems and the capstone project.

Participants in this course learn how to analyze course-related datasets, calculate cosine similarity, and build a course similarity matrix. Additionally, they will generate various types of recommendation systems by performing KNN-based collaborative filtering, PCA, and non-negative matrix factorization.

Throughout this course, participants will gain hands-on experience in building recommender systems through various projects. Finally, they will share their work with peers and evaluate each other’s performance.

Meet the Instructor

Yan Lou – Ph.D., Data Scientist and Developer

Yan Lou is currently a senior machine learning engineer at Criteo, Canada. Previously, he worked as a data scientist and developer at IBM. In 2016, he received his Ph.D. in machine learning from the University of Western Ontario.  Over the years, Yan Lou has built multiple AI and cognitive applications in digital banking, mining software repositories, and personal healthcare management.

Artem Arutyunov – Data Scientist

Artem Arutyunov is a data scientist on the skills network team at IBM Canada. He has also authored multiple machine learning courses and projects. He did his bachelor of science, mathematics, and statistics from the University of Toronto.